So why not both? A glow bar could turn on when the oven was turned on, thereby lighting a pilot light, which would then stay on as long as the oven was on. Ah, but a rhetorical question, since the issue with pilot lights is that they can blow out, and if the gas valve turns on without a pilot (or glow bar) to re-ignite the oven, the house fills with gas and any little spark will ignite far more than the gas burner.

So there’s a safety device on pilot-lit gas values called a thermocouple — the heat of the pilot light causes a length of copper wire to expand, pushing open the safety valve — pilot out, safety valve closes and … well the good news is that your house doesn’t explode; the bad news is that this tends to happen on Saturday so you call a plumber to replace this $5 part at double-time and $155 later, you’re good to go.

But the pilot + thermocouple means the pilot’s on all the time. A glow bar uses a separate safety mechanism (I am guessing) — if the glow bar burns out, there’s no current, and without current, the valve won’t open. Same result, different solution. So some genius said “the glow bar will only be on when the oven is being used and but the gas is off because the oven is warm enough”. Smart genius, lousy engineer created a thing that uses 300W to light up a very thick bit of metal enough to cause it to ignite gas — has to be thick because it would melt under the heat of the regular oven if it were thinner. And, as a plumber who commented with the original answer noted, the glow bar eventually burns out (like the thermocouple) and needs to be replaced, giving a nice source of incremental revenue to the service companies that fix these things for us.

Of course they could come up with a simple solution that would a) use little or no electricity, b) use little or no extra gas, c) be reliable and safe, and d) last for the lifetime of the oven. But that would be complicated … and also would cut in to those high-markup service calls. Am I cynical, or is that all there is to this?

I’ve had some problems with the TED system I am trying to get resolved, but over measurement isn’t one of them. Can you point me to some hits at what you have done to isolate consumption?

Energy monitoring is cool because unlike the weather, you have some ability to respond and take control. We use much less electricity than we used to because by keeping track of our use, we have found numerous ways to keep the house more comfortable and properly lit while using considerably less energy. We’re saving around $100 per month on our electricity bill as a result of understanding all the little things that use electricity without knowing.

My baseload of 75 watts didn’t come easy. It took a lot of time and energy to track down every little always-on device in the house and make a decision about whether we really needed it. In the end, the biggest difference was when we got rid of cable TV and the set-top box that chewed up many watts even in standby mode. By the way, we don’t even miss it.

What’s left are a cable modem box (for internet access), a router, cordless phones, and a few clocks.

In all honesty, the only reason we can get this low is that only my wife and I live here now. If our children were still with us, we’d be hardpressed to hit 250 watts baseload.

David

The energy graphs at http://energy.shadypixel.com are made from a bunch of open source software that I strung together (under the guidance of my son Avery Fay, who is the one that really understands all this software stuff).

It works like this. The software framework is collectd (http://collectd.org/) — a program that periodically collects information from computers or other hardware using input modules that have been written by various contributors. Collectd then passes the information along to output modules which store the information in a database, pass it to another computer, graph it and so forth.

The input module I use (“ted” by name) reads the household power levels from The Energy Detective (TED). I have it set to read TED every second (the fastest TED can be read), giving me the highest resolution energy traces I have seen on the web.

The output module I use is called rrdtool, for round-robin database tool (http://oss.oetiker.ch/rrdtool/). It passes the TED energy values to another computer that is on the web (shadypixel.com), which then puts them into a round-robin database (so no database management is required) and then creates graphs from the data which I use to update a web page (energy.shadypixel.com) every minute.

But maybe the coolest thing (literally) is that the computer that runs collectd is a plug computer (SheevaPlug running linux) that I bought for $100 (its projected cost long term is $50) and which consumes only 5 watts. This is a big improvement over my former setup in which collectd ran on my laptop, which uses 25 watts and which had to be disconnected from TED whenever I needed to take it on the road. With a SheevaPlug dedicated to TED, I get a 24 x 7 energy record from anywhere on the Web and I don’t feel guilty about burning 5 watts to get it.

David

I’ve been thinking along similar lines — writing software to identify the energy signature of each appliance. This approach even has a name: Non-Intrusive Load Monitoring. Here’s a website with some background on NILM and its history: http://www.openamr.org/wiki/NILM

NILM is an active academic research area but I think a different approach may be more productive — multichannel residential energy monitors. Imagine having a separate monitor for each circuit in your house. There would still be multiple light fixtures, for example, on each circuit, but the heavy hitter energy consuming appliances tend to be on separate circuits so you could get a separate trace for each one. At least I think so. I haven’t tried this yet, although I’ve identified a good candidate for the monitor: the Brultech ECM-1240. It costs about the same as a TED but has 7 channels so it can monitor total energy plus six circuits. What’s more, you can chain multiple ECM-1240′s together. Very exciting and the hardware hackers at cocoontech.com are all over it (http://www.cocoontech.com/forums/index.php?showtopic=12187&st=135)

Ultimately, we would like to have smart devices that monitor and report their energy usage over house wiring to a central server. But I’m not holding my breath. In the meantime, monitoring at the circuit level may be the next step.

David